INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Meningococcal disease is a significant cause of mortality and morbidity throughout the world. Neisseria meningitidis strains of serogroup B are the most common cause of sporadic meningococcal diseases in developed countries (31). Within serogroup B, most disease is caused by a limited number of groups of genetically related bacteria that have been referred to as complexes, clusters, or lineages. These are the ET-5 complex, the ET-37 complex, lineage III, and cluster A4 (7). These clonal groups have been associated with an increased rate of disease and reinforce the need for a serogroup B vaccine. A serogroup B polysaccharide-based vaccine is not yet available. Some issues related to its structure identity with host cell molecules, such as neural cell adhesion molecule (13), has led research to focus on other bacterial components. Multilocus enzyme electrophoresis studies have shown that lineages of serogroup B meningococci diversify during spread and that their members often change antigenically (28, 30, 46), raising real concerns for the development of a vaccine (46) and reinforcing the need to study antigenic and genetic variation of vaccine target antigens among members of these lineages.

Among antigens considered for inclusion in a future meningococcal vaccine is the human transferrin receptor, which is composed of two subunits, TbpA and TbpB (43). TbpB has several attributes of a good vaccine candidate: it is a surface-exposed molecule, expressed in vivo during infection (1, 2, 3, 6, 16, 23), and it elicits protective and bactericidal antibodies in laboratory animals (3, 10, 24) and opsonic antibodies in humans (23). Moreover, a phase I clinical trial has shown that a recombinant TbpB (rTbpB) purified from Escherichia coli was safe and immunogenic in humans (B. Danve, F. Guinet, E. Boutry, D. Speck, M. Cadoz, L. Lissolo, X. Nassif, and M. J. Quentin-Millet, presented at the 11th Int. Pathogenic Neisseria Conf., 1998). Besides the meningococcal TbpB, native TbpB and rTbpB from other bacterial pathogens were shown to confer protection or induce bactericidal antibodies against the corresponding homologous strain. This protective role of the antibodies has been shown with native TbpB (initially named TfbA) from Actinobacillus pleuropneumoniae (41) or rTbpB from Haemophilus influenzae (25), Moraxella catarrhalis (29) and, more recently, Pasteurella haemolytica (34).

The variability observed in the TbpB protein from N. meningitidis (12, 17) has raised questions about its capacity to be used as a broadly cross-reactive antigen. However, we have shown that while TbpB varied among strains, antigenic and genomic features of TbpB and tbpB allowed the meningococcal strains to be classified into two major families: isotype I (tbpB gene of 1.8 kb and TbpB protein with a mass of approximately 68 kDa) or isotype II (tbpB gene of 2.1 kb and TbpB protein with a mass of approximately 80 to 90 kDa) (38).

Protection against meningococcal disease has been correlated with the persistence of bactericidal antibodies, and various lines of evidence have highlighted the importance of humoral bactericidal activity in host defense against N. meningitidis (14). In this context, several studies on Tbp molecules have investigated the induction of cross-bactericidal antibodies, and the specific role of TbpB in regard to this induction was investigated (1, 10, 15, 24, 29, 34, 39, 40). Our previous work with rTbpB molecules purified from strain M982 (representative of isotype II strains) showed that it was possible to induce cross-bactericidal antibodies against isotype II strains, but no cross-bactericidal activity was seen with strain B16B6, a representative of isotype I (40). However, to date, total cross-bactericidal activity has not been reached with any anti-TbpB or rTbpB tested. Major TbpB variants have been identified among isotype II strains, suggesting that more than one TbpB protein may be required to raise fully cross-bactericidal sera within this isotype (39, 40). The variability of tbpB gene and the TbpB protein has not yet been assessed on a representative collection of serogroup B strains. The distribution of the two isotypes, I and II, is not known among strains belonging to the different clonal complexes implicated in serogroup B meningococcal disease.

The aim of the present study was to assess the distribution of the two isotypes of tbpB among serogroup B strains belonging to the four major clonal complexes responsible for meningococcal B disease and to characterize the extent of the allelic diversity of the two isotypes. A collection of 108 isolates was analyzed by PCR restriction fragment length polymorphism and analysis and sequencing of representative tbpB genes. A panel of 27 strains representative of the diversity of the tbpB gene was then selected to assess the number of rTbpB variants necessary to induce the broadest possible cross-bactericidal activity. Moreover, using rTbpB produced in fusion with the maltose-binding protein (MBP) corresponding to different parts of the TbpB molecule, we investigated which part of the TbpB molecule was involved in the induction of cross-reacting antibodies.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, culture conditions, and E. coli constructions. The 108 meningococcal isolates examined here are listed in Table 1. The collection included 30 strains belonging to the ET-5 complex previously described (39), 29 strains belonging to lineage III (the 17 isolates from New Zealand were kindly provided by D. R. Martin), 27 strains belong to cluster A4, and 22 isolates of the ET-37 complex. Five reference strains belonging to other clonal complexes, B16B6, M982, M978, S3032, and 6940 (21, 27), were also included. The isolates were characterized by their combination of alleles at 14 enzyme loci as described previously (8) and then serotyped and subtyped with monoclonal antibodies (33). For DNA extraction, N. meningitidis strains were grown on Mueller-Hinton agar (MHA) plates (Difco).

E. coli constructs. Antibiotic-resistant E. coli bacteria were selected on Luria broth plates containing 100 µg of ampicillin per ml. The different MBP translational fusions of TbpB used in this study are described in Table 2: the fusion proteins corresponding to the N terminus and the C terminus from N. meningitidis BZ83 were obtained after PCR amplification using primers 5'-CGGGATCCCTGGGCGGAGGCGGCAG-3' and 5'-AAAACTGCAGATTATTCCAGTTTGTCTTTGGTTTTCG-3' for the N-terminus construct and primers 5'-CGGGATCCAATGGCGCGGCGGCTTCAG-3' and 5'-AAAACTGCAGATTATTGCACAGGCTTTTGGCG-3' for the C-terminus construct. Convenient restriction sites were introduced (italicized letters) in order to clone these fragments in pMal-C2 (BamHI and PstI).

Characterization of tbpB genes by PCR amplification and restriction analysis of the amplified fragment. N. meningitidis strains were grown on MHA plates (Difco). Extraction of DNA from each strain was performed with a rapid method using guanidium isothiocyanate (32). PCR amplifications of the tbpB genes were performed on genomic DNA using two primers, OTG 6687 and OTG 6689, described by Legrain et al. (22), under the following conditions: 25 cycles of denaturation of DNA at 94°C for 1 min, with annealing at 58°C for 2 min, and with extension at 72°C for 3 min. Each PCR was analyzed on 1% agarose gel, and the size of the tbpB gene was determined using a molecular weight marker, HindIII (New England Biolabs). The amplified fragment was purified on a Qiaquick column (Qiagen) and then digested by using the enzymes HincII, AvaII, VspI, XhoI, and/or BamHI and MluI in six separate reactions according to the protocols specified by the manufacturer (New England Biolabs). The restriction products were separated by electrophoresis on 2% agarose gels. The gels were stained with ethidium bromide and photographed, and the patterns were visually compared. A second PCR was carried out on genomic DNA primers P3 and P4 (39). Reactions were performed in a 100-µl volume containing 200 µM concentrations (each) of dCTP, dGTP, dATP, and dTTP (Pharmacia-LKB); 0.2 µM concentrations of each primer, and 2.5 U of Taq polymerase (Appligene). Amplifications were performed in a DNA thermocycler (Biometra; Trio-Thermobloc) programmed as follows: initial denaturation at 95°C for 5 min, followed by 25 cycles of consecutive denaturation (30 s, 95°C), annealing (30 s at 58°C), and DNA chain extension (1 min at 72°C). The size of the amplified fragment was determined after electrophoresis of PCR product on a 3% agarose gel.

Characterization of new tbpB genes by sequence analysis. Fragments obtained after PCR with primers OTG6687 and OTG6689) were cloned in a pCR-Topo vector (Invitrogen) and automated cyclo-sequenced (Genome Express, Grenoble, France). When results were questionable or unexpected, the nucleotide sequence was confirmed by sequencing another clone obtained during another PCR. Nucleotide and deduced amino acid sequences were compared using the CLUSTAL W multialignment software from the Infobiogen package. A phylogenetic tree was constructed by the neighbor-joining method of Saitou and Nei (42) using the Grow Tree software from the Genetics Computer Group (GCG) package.

Purification of meningococcal Tbp complexes, TbpB subunits, and recombinant TbpB or MBP-TbpB fusions. The TbpA+B complexes were prepared as previously described from N. meningitidis B16B6 (B:2a:P1.2 with a TbpB with a mass of ca. 68 kDa), M982 (B:9:P1.9 with a TbpB with a mass of ca. 88 kDa), 8680 (B:15:P1.3 with a TbpB with a mass of ca. 82 kDa), and BZ83 (B:15: with a TbpB with a mass of ca. 81 kDa) (10). Strains B16B6 and M982 were chosen as representative of the two major isotypes, I and II (38). Strains BZ83 and 8680 were chosen as representative of the two major variants of isotype II (39). Purified TbpB from strains B16B6 and M982 were obtained as previously described (24). The MBP-TbpB fusion molecules were produced by induction of recombinant E. coli cultures at log phase for 3 h with 0.3 mM isopropyl--D-thiogalactopyranoside (IPTG; Appligene) and purified by using an amylose affinity column (36). The nonfusion rTbpB of strain B16B6 was purified from the membrane fraction of E. coli after cell disruption and centrifugation. Briefly, the membrane fraction was solubilized in buffer containing Zwittergent 3-14 (Calbiochem) and was chromatographed on Q-Sepharose columns. Fractions containing rTbpB were concentrated and loaded on an S-300 column for final polishing to remove endotoxins and degradation products. The purified protein was renatured and stored at <45°C.

Preparation and characterization of mouse and rabbit antisera. New Zealand White rabbits were immunized by the subcutaneous and intramuscular routes with three injections of native TbpB protein from strain B16B6 or TbpA+B complex from the four meningococcal strains (B16B6, M982, 8680, and BZ83) or with the eight MBP fusions. Briefly, 100 µg of protein were emulsified in complete Freund adjuvant and injected on days 0, 21, and 42. On day 57, the animals were sacrificed and the blood was collected. For bactericidal assays, rabbit immunoglobulin G (IgG) antibodies were affinity purified on protein G-Sepharose 4FF columns (Pharmacia). The protein concentration of each IgG preparation was determined by a MicroBCA assay (Pierce).

Serum bactericidal assay. The bactericidal activity of purified rabbit IgG and mouse sera was tested as described earlier with slight modifications (40). Briefly, 50 µl of twofold serial dilutions of IgG solutions or serum were added to 96-well microtiter plates (Nunc) and incubated with 25 µl of an iron-starved meningococci suspension adjusted to 2 × 10⁴ CFU/ml and 25 µl of baby rabbit complement. After 1 h of incubation at 37°C, 20 µl of the mixture from each well was plated onto MHA plates. The plates were incubated overnight at 37°C in 10% CO₂. The bactericidal titer of each serum or IgG preparation was expressed as the inverse of the last dilution of serum at which 50% killing was observed compared to the complement control. For complement-sensitive strains, the complement was adsorbed on formaldehyde-fixed bacteria (10¹⁰ CFU [3 ml of complement]) before being used.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Characterization of tbpB gene from serogroup B strains. The 108 N. meningitidis strains of serogroup B analyzed were assigned to four major epidemiological complexes: ET-37 complex (22 strains); lineage III (29 strains), cluster A4 (27 strains), and ET-5 complex (30 strains). They corresponded to 6 different serotypes and 15 different serosubtypes and were recovered from 21 countries in various part of the world between 1966 and 1996 (Table 1). After PCR amplification with primers OTG6687 and OTG6689 of the tbpB gene of each isolate, a unique PCR product with an apparent size of approximately 1.8 kb (isotype I) or 2.1 kb (isotype II) was obtained (Table 1). Isotype I was exclusively found among the ET-37 complex strains. Of the 22 strains of the ET-37 complex strain, 1 displayed a tbpB gene corresponding to isotype II. In the three other clone complexes, all strains had a tbpB gene corresponding to isotype II (2.1 kb). Overall, 80.6% of the strains analyzed corresponded to isotype II and 19.4% were of isotype I.

Sequence analyses of tbpB genes from serogroup B strains. In order to precisely determine the degree of heterogeneity of the tbpB alleles of serogroup B strains corresponding to the two isotypes, the complete nucleotide sequences of tbpB of 10 new serogroup B strains were determined. The sequences were compared to each other and to all other known sequences of serogroup B tbpB alleles by multialignment using the CLUSTAL W program. In all, 24 full-length tbpB sequences (corresponding to 5 isotype I strains and 19 isotype II strains) were compared. The percentage of nucleotide divergence among isotype I strains varied from 1 to 33% (1 to 42% for the deduced amino acid sequences). Among isotype II strains, the level of heterogeneity of tbpB nucleotide sequences varied between 1 to 24% (1 to 33% for the corresponding amino acid sequences). The three major variant types among the isotype II strains were confirmed by nucleotide sequence analysis. Strains belonging to the BZ83-like and 8680-like groups presented a low level of heterogeneity of tbpB nucleotide sequences. The BZ83-like strains differed from one another at 1 to 4% of their tbpB nucleotide sequences, confirming the PCR typing results and providing additional evidence that the BZ83 subtype is well conserved even among strains belonging to different complexes.

View larger version (24K): [in this window] [in a new window]   FIG. 1.   Phylogenetic tree of tbpB nucleotide sequences of 24 strains of different clonal groups and geographic origins. The name of the strain, its tbpB allele, its geographic origin, and its clonal group are indicated, as is the year of its isolation. The prototype strain for each major variant is indicated in boldface. The phylogenetic tree was constructed by the neighbor-joining method of Saitou and Nei (42). The EMBL accession numbers of the 10 new tbpB nucleotide sequences are AJ279554 (strain 2717), AJ279555 (strain 901256), AJ279556 (strain 64/92), AJ279557 (strain 90/94), AJ279558 (strain 92123), AJ279559 (strain M101/93), AJ279560 (strain AO21), AJ279561 (strain M2), and AJ279563 (SB22). The names and sources of other tbpB nucleotide sequences referred to in this study are strains B16B6 and M982 (21), strains BZ163 and BZ83 (20), strains 106, 2713, and NGPB24 (22), strains 6940, M978, and S3032 (27), and strains 8680, 8710, 32/94 and 8726 (39).

Antigenic relevance of the four major genetic tbpB variants. The antigenic relevance of the four major variants was investigated. For this purpose, rabbit sera raised against the MBP fusion of the full-length TbpB from four meningococcal strains were produced (Table 2). Their cross-reactivities against the TbpB protein of the four strains B16B6, M982, BZ83, and 8680 were assessed by Western blotting, and their bactericidal activities against the four strains were determined (Table 4). The anti-MBP control sera did not react with any of the strains as expected. The four rTbpB fusions were shown to be immunogenic and induced TbpB-specific antibodies. All individual sera specifically recognized their homologous TbpB on the basis of a Western blotting on purified Tbp complex. TbpB-specific IgG purified from these sera displayed a bactericidal activity against their homologous strain. Cross-reactivity was detected for all sera when tested in a Western blot against the TbpB of the four strains. However, no cross-bactericidal activity was detected with an IgG preparation raised against an isotype I strain (B16B6) or isotype II strains (M982, BZ83, and 8680). Reciprocally, sera raised against rTbpB fusions from the three isotype II variants were not bactericidal against the isotype I strain B16B6. Sera raised against the rTbpB from isotype II strains displayed full cross-reactivity against the three isotype II strains on the basis of Western blot analysis. However, their cross-bactericidal activity toward the three nonhomologous strains was variable. For example, IgG raised to rTbpB-M982 killed preferentially strain BZ83 (titer of 64), while IgG raised to rTbpB-BZ83 killed only M982 (titer of 16). IgG raised to rTbpB from strain 8680 killed only the homologous strain, and an incomplete killing was observed with the other ET-5 complex strain (BZ83). In all cases, the level of killing was higher with the homologous strain. These results clearly showed that cross-reactivity of sera based on Western blot analysis was not predictive of their cross-bactericidal activity. Moreover, the M982 variant seemed to be antigenically closer to BZ83 than 8680. 

Contribution of the N-terminal and C-terminal domains of TbpB in the induction of cross-reacting antibodies. In an attempt to identify which part of the TbpB molecule was able to induce cross-reacting antibodies, rabbit sera were raised against the N-terminal and C-terminal domains of TbpB from strains M982 and BZ83 produced in fusion with MBP (Table 2). The corresponding IgG preparations were analyzed for their ability to induce cross-reacting antibodies by Western blot and bactericidal assays (Table 5). The results show that both TbpB domains from these two strains were immunogenic by themselves. As already observed with the anti-MBP-TbpB (FL), both domains induced antibodies that were cross-reactive in the Western blot assay with the heterologous TbpB from the Tbp complex purified from the three isotype II strains, even with the TbpB from isotype I strain B16B6. Again, however, no bactericidal activity was detected against this latter strain. Antibodies raised against the C-terminal domain of TbpB of strain M982 and strain BZ83 displayed a higher bactericidal activity than those raised against the corresponding N-terminal part of the protein. The antibodies raised to the C-terminal domain of TbpB from strain M982 were able to kill the strain BZ83, but the reverse was not observed. These results showed that both N- and C-terminal domains individually expressed from variant M982 in fusion with MBP presented epitopes commonly accessible on the surface of strains M982 and BZ83.

Complement-mediated bactericidal activity of specific antisera against a panel of 27 meningococcal strains. A panel of 27 strains was chosen to represent the diversity observed in tbpB alleles. These strains are indicated by an asterisk in Table 1. The four major variants, B16B6, M982, 8680, and BZ83, and similar strains were all represented in the selected panel in the same proportion as in the complete collection (chi-square test, ² = 1.76, P = 0.62). All strains were able to bind horseradish peroxidase-hTf and expressed TbpB, and all of the strains can be killed by a rabbit antiserum raised against meningococcus whole cells, indicating that all of the tested strains were not resistant to killing (data not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Ever since the identification of the two isotypes of the tbpB gene (38), their distribution among strains of N. meningitidis responsible for meningococcal disease has remained unclear. In this study, the distribution of the two tbpB isotypes was determined by analyzing 108 strains of the four major clonal complexes responsible for serogroup B disease. Isotype I strains were less represented than isotype II strains (19.4 versus 80.6%). This distribution is specific to the present collection corresponding exclusively to serogroup B strains, and one could expect that the proportions of the two isotypes would be different if serogroup A or C strains were included.

The extent of the diversity within the two isotypes was characterized, and 14 different alleles were identified based on restriction analysis of the tbpB gene. We demonstrated here that variability of tbpB was not restricted to isotype II, in contrast to what was thought previously on the basis of a study of a limited number of serogroup B strains (39). Among isotype I strains, the prevalent tbpB allele was a-12 and corresponded to the prototype strain B16B6, but two other alleles were identified (a-13 and a-14) for which the levels of nucleotide homology compared to a-12 were 58 and 75%, respectively. Despite this nucleotide sequence variability, antibodies raised against different preparations of TbpB from the prototype strain B16B6 showed cross-bactericidal activity (Table 6). In the present study, isotype I appears to be strictly associated with the ET-37 complex. This result is in agreement with the study of Brieske et al., who showed that all of the 19 ET-37 complex serogroup C strains studied belonged to isotype I, but in that study no variation of tbpB gene was detected (6). Despite the fact that the ET-37 complex has been described as a quite-homogeneous complex that comprises a limited number of variant clones (9, 45), it has been shown that strains from this complex were extremely variable with regard to their Opa proteins; 26 Opa proteins were identified by using monoclonal antibodies (45). This variation was described as resulting from recombinational reassortment, including duplication and import by horizontal genetic exchange (18). In the case of the tbpB allele, it is likely that recombination and/or horizontal genetic exchange events led to the appearance of the new variants described. Such events have been suggested in two previous studies of the tbpB gene of isotype II (20, 22). Horizontal and genetic exchange could result in identical or almost identical genes in unrelated meningococci; for example, strain 901256 from The Netherlands, which belongs to ET-37 complex, had the same tbpB allele as strain BZ163, also from The Netherlands, which belongs to cluster A4. Such an observation illustrates the concept of the global gene pool described by Maiden et al. (26).

This study demonstrates that the three major nucleotide tbpB gene variants (M982, BZ83, and 8680) initially detected among ET-5 complex strains (39) also exist among strains of isotype II belonging to other clonal complexes. Type M982-tbpB variants were largely represented among cluster A4 and lineage III strains and were variable at the nucleotide sequence level. Type BZ83-tbpB variants were preferentially represented among ET-5 complex strains but were also found in cluster A4 and lineage III and seemed to be very conserved. 8680-tbpB variants were not restricted to the ET-5 complex and were also found in cluster A4. The relevance of these variants was also shown on the basis of the ability of sera raised to the corresponding rTbpB to induce cross-bactericidal antibodies (Tables 4 and 5), since the number and the identities of the strains killed by the four anti-rTbpB sera were not the same (Table 6).

The results presented here provide further evidence that a recombinant TbpB is able to induce cross-bactericidal antibodies. This was previously shown with rTbpB from strain M982 against a smaller collection of isotype II strains (40). The present study allows us to extend the observation to isotype I strains by using rTbpB from strain B16B6. We also showed that the C-terminal part of the molecule was able to induce by itself cross-bactericidal antibodies against a large panel of strains, in addition to its ability to induce bactericidal antibodies against the homologous strain that has been previously reported (36). The TbpB C-terminal domain corresponds to the most conserved part of TbpB molecules (according to multialignment analysis performed with the CLUSTAL program, data not shown), and common C-terminal epitopes are most likely surface exposed on N. meningitidis strains.

This study highlights the caution with which cross-reacting antibodies need to be studied and analyzed; in this work, both Western blot and bactericidal assays were used, and they led to different results. It has been shown that antibodies to meningococcal Tbp complexes and to isolated proteins have been detected in both patients and carriers and that these antibodies were able to cross-react with TbpA or TbpB isolated from different meningococci (3, 11, 16, 19). However, in all cases, these conclusions were based on Western blot analysis or enzyme-linked immunosorbent assay (ELISA) performed on purified Tbp complexes. Under these conditions, the cross-reactivity observed with antibodies by these techniques could not be predictive of their cross-functional activity. This is clearly shown in the present study, where cross-reactivity between the two proteins of the Tbp complex from isotype I strains and from isotype II strains was detected by Western blotting with specific antiserum, but antibodies raised to one or the other Tbp complex failed to induce bactericidal activity against the heterologous isotype strain. Using a different technique, Lehmann et al. showed that human serum from an N. meningitidis-infected patient was able to induce cross-opsonins against Tbp complex independently of the isotype of the strains. This conclusion was based on the use of beads coated with purified Tbp complex corresponding to the two isotypes, and under such conditions the epitopes may be distinct or may present different accessibilities than those exposed on live, viable meningococci (23).

The extent of cross-bactericidal activity observed with antisera raised to the rTbpB molecules was slightly lower than that observed with antisera raised to the TbpA+B complex (Table 6). At least three nonmutually exclusive hypotheses may explain this result: (i) TbpB retains a more native conformation when it is present within the complex rather than isolated, (ii) the presence of hTf in the purification scheme may help TbpB to adopt a conformation closer to the native one, or (iii) anti-TbpA antibodies directly contribute to the bactericidal activity of the antiserum. That hTf contact may help TbpB in refolding is indirectly supported by several studies which demonstrated that meningococcal transferrin binding is determined by independent specific recognition sites in hTf localized primarily in the C lobe but also in the N lobe (4, 5, 36, 37). The results of killing scores of isotype I strains with antisera to native TbpB from strain B16B6 purified on hTf (five strains killed out of five total) compared to the score obtained with the corresponding MBP fusion (three of five strains) (Table 6) may confirm this hypothesis. However, for pharmaceutical considerations, we chose not to purify the clinical-grade rTbpB on an hTf affinity column, and we developed an alternative scheme which gave intermediate results (four of five strains killed). The role of TbpA in the induction of protective antibodies is difficult to define, since no functional studies with separate TbpA and TbpB molecules as target antigens are available to date to give a definitive answer on the role of TbpA of N. meningitidis. However, a study of the protective capacity of P. haemolytica TbpA and TbpB in cattle showed that the best protection was observed in the group of animals immunized with both TbpA and TbpB. There was a high antibody response against TbpB and not against TbpA; no protection was detected in the group immunized with TbpA alone, and no antibody-mediated response was observed in this group. This led the authors of that study to suggest that TbpA contributed to protection through the induction of a non-antibody-mediated response (34).

Among isotype II strains, no single TbpB molecule was able to induce an antiserum displaying complete cross-bactericidal activity against all the strains. At least two of the three tbpB variants (M982 and 8680) seemed to be required to extend the level of cross-bactericidal activity, since different strains were killed depending on the variant of TbpB used to induce the antibodies.

Overall, using a large panel of strains, we demonstrated here that the two isotypes of tbpB are not equally represented among strains representative of serogroup B disease. We showed that four major variants of tbpB gene characterized at the nucleotide level are represented in the collection. The rTbpB proteins corresponding to these four variants were able to induce cross-bactericidal antibodies reacting with different strains from the panel studied. Since the data presented here were obtained with a collection of strains deliberately chosen to represent the largest possible diversity of TbpB and were carried out under stringent conditions, only four major variants were found. With this encouraging finding, we believe that an efficient, broadly cross-reactive rTbpB-based vaccine is feasible. Based on the results presented here, this vaccine should include at least two TbpB variants (B16B6 and M982); the addition of a third variant, if necessary, should preferably be variant 8680. Formulations and adjuvants are important aspects which will impact the immunogenicity of rTbpB. One could expect that these would modify the immunogenicity and potentially increase the cross-reactivity of rTbpB molecules. Our laboratory is currently investigating these issues in order to define the exact number of rTbpB molecules in the final product.